AI agents call graph_query to retrieve information from Mementos without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool performs a read-only traversal of a knowledge graph structure, returning existing entities and relations. There are no side effects, no data modifications, and no external code execution. This is a straightforward retrieval operation similar to a search or fetch function. Severity is low because the blast radius of misuse is limited to information disclosure within the agent's own memory system.
From the tool's definition Tool description states 'Traverse the knowledge graph from an entity up to N hops. Returns entities and relations.' This is a query operation that retrieves and returns data without modifying, deleting, or executing external actions.
Attacks that exploit this kind of access
Traverse the knowledge graph from an entity up to N hops. Returns entities and relations. It is categorised as a Read tool in the Mementos MCP Server, which means it retrieves data without modifying state.
Register the Mementos MCP server in PolicyLayer and add a rule for graph_query: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Mementos. Nothing to install.
graph_query is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the graph_query rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for graph_query. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
graph_query is provided by the Mementos MCP server (@hasna/mementos). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →